When cold-starting a "full-FATES" run, we two have ways of doing this (once #1558 is integrated). Whether we use an abundance type initialization, or a DBH type, both assume that every grid-cell contains all PFTs defined in our parameter file. However, I think there are other ways to approach this.
One alternative would be to use the surface file information to initialize a full-FATES run, in the same way we initialize a fixed-biogeography run, however we wouldn't maintain the PFT restrictions on the grid-cells the way we do with fixed-biogeography. One universal benefit is that it incorporates permanently bare-ground patches into the simulation.
It could be run with supplementary seed rain to ensure biodiversity.
It also allows us another intermediate form of comparison with fixed biogeography runs, which would help us see the drift from fixed-biogeo initial state more clearly.
I don't think this is an urgent need seeing how the upcoming FATES release will focus on nocomp runs, but it would be nice to bat some of these ideas around so we have a roadmap for future releases.
When cold-starting a "full-FATES" run, we two have ways of doing this (once #1558 is integrated). Whether we use an abundance type initialization, or a DBH type, both assume that every grid-cell contains all PFTs defined in our parameter file. However, I think there are other ways to approach this.
One alternative would be to use the surface file information to initialize a full-FATES run, in the same way we initialize a fixed-biogeography run, however we wouldn't maintain the PFT restrictions on the grid-cells the way we do with fixed-biogeography. One universal benefit is that it incorporates permanently bare-ground patches into the simulation.
It could be run with supplementary seed rain to ensure biodiversity.
It also allows us another intermediate form of comparison with fixed biogeography runs, which would help us see the drift from fixed-biogeo initial state more clearly.
I don't think this is an urgent need seeing how the upcoming FATES release will focus on nocomp runs, but it would be nice to bat some of these ideas around so we have a roadmap for future releases.